The package xxl.core.math.statistics.parametric.aggregates provides implementations of the techniques proposed in [HHW97]: J. Hellerstein, P. Haas, and H. Wang, 1997. There, parametric estimators are computed online over a data stream and in some cases they are even equipped with the according confidence intervals. The online computation of the aggregation functions relies on the {@link xxl.core.cursors.mappers.Aggregator aggregator cursor}. An aggregate is computed by an aggregation function inductively with agg_n = f (agg_n-1, next). This means an aggregation function must be a two-figured function expecting as first argument the previous computed aggregation value and as second argument the next object to use for computation.